Antibodies to inner ear antigens in Meniere's disease

Abstract

Meniere's disease (MD) is an idiopathic inner ear disorder characterized by fluctuating hearing loss, episodic vertigo and tinnitus. Its aetiology is unknown, although there is growing evidence that autoimmunity may be involved in its development. Using the Western blot immunoassay, we examined the reactivity to bovine inner ear antigens of sera from a series of MD patients who had previously been extensively studied for the presence of antibodies to collagens and membrane proteins. Reactivity to inner ear antigens of molecular weight 44 and 53 kD was found in 11/25 (44%) and 10/25 (40%) of the patients, respectively; both antigens were absent in the sera of healthy donors. It is still unclear whether the antibodies to 44 and 53 kD proteins play a role in the pathogenesis of MD or if they instead represent the result of inflammation and tissue destruction. Even if the latter is true, they may contribute to the perpetuation of the disease or play a role as a cofactor in association with other mechanisms.

INTRODUCTION

Meniere's disease (MD) is an idiopathic inner ear disorder characterized by fluctuating hearing loss, episodic vertigo and tinnitus in which the histopathological lesion takes the form of endolymphatic hydrops. Vascular disturbances or infectious agents may be involved in its development, but the etiopathogenesis of the disease is still unknown. In about one-third of all cases the disease seems to be of autoimmune origin, although the immunological mechanisms involved are not yet clear [1,2]. Recently various groups have focused their attention on the immune response to inner ear antigens and a number of studies have sought to identify these autoantigens. Wei et al. [3] found antibodies reactive with autologous ganglion cells in MD patients, but not in patients with temporal bone trauma. The authors suggested that alterations in the microvascolature due to an autoimmune reaction in the spiral ganglion could trigger MD. Raised levels of anti-collagen II antibodies have also been reported in a high percentage of MD patients by Yoo et al. [4–6], although these findings were not confirmed by others. In 1994 we analysed the levels of antibodies to collagen (C) II, V, I and basement membrane proteins (laminin and CIV) in MD patients, but a clear role of these autoantibodies in the pathogenesis of MD could not be established [7]. Thus, despite many studies the role of antibodies reactive with structures of the inner ear in the pathogenesis of MD remains unsettled.

The aim of the present study was to evaluate the reactivity of sera from MD patients with inner ear antigens, and to correlate the serological findings with the clinical manifestations of the disease.

PATIENTS AND METHODS

Patients

Twenty-five patients affected by Meniere's disease were studied: 11 women and 14 men, aged 32–76 years (mean age ± SD: 51·16 ± 13·8 years). In all patients the disease was monolateral with a duration of more than 9 years (mean duration ± SD: 11·96 ± 2·2 years).

The diagnosis of MD was made based on the criteria contained in the 1995 Committee on Hearing and Equilibrium Guidelines [8]. The patients were given a thorough audiological examination, including pure tone audiometry, clinical timpanometry, and brain stem-evoked response audiometry tests. Vestibular function was assessed by videonystagmography (VNG Ulmer 1·3 System) to detect the presence of spontaneous nystagmus, positional and positioning nystagmus, and head-shaking nystagmus. The Fitzerald-Hallpike caloric tests were also performed. In all patients the investigations were carried out in intercritical periods. The patients also underwent brain computerized axial tomography and/or magnetic nuclear resonance scans.

Pure tone audiometry showed sensorineural low frequency hearing loss (LFHL) in 11 (44%) patients and sensorineural flat hearing loss (FHL) in 14 (56%). The level of hearing loss was assessed by four frequency pure-tone averages (PTA) ranging from 500 to 3000 Hz. No cases of high frequency hearing loss were disclosed. The Auditory Evoked Brainstem Responses test (AEBR) confirmed the cochlear origin of the hearing loss.

Videonystagmography showed positional nystagmus in 7 patients. Head-shaking nystagmus (HSN) was positive in 11 cases and the caloric test revealed canal paresis in all patients. After their informed consent was obtained, serum samples were taken and immediately stored at −20°C. Sera from 38 healthy donors (18 women, 20 men, age 25–75, mean age ± SD: 40 ± 12 years) were used as controls. No one suffered from otologic disorder or systemic autoimmune diseases.

All sera were tested for antibodies to inner ear, spleen and liver antigens and for non-organ specific antigens.

Isolation of inner ear

Inner ear material was isolated from fresh bovine temporal bones obtained from a local abattoir. Temporal bone microdissection was carried out using a median approach following the petrous ridge and internal auditory canal. After the dura was removed from the temporal bone, the superior portion of the internal auditory canal was drilled, and the basal turn of the cochlea was sought for immediately beneath the first genu of the facial nerve. Continuing the anterolateral dissection, the cochlea was totally exposed; the vestibule was found by drilling behind the internal auditory canal. The membranous structures of the inner ear were opened by a diamond burr and extracted with a hook and cup forceps [9].

Tissue extracts

The dissected bovine inner ear was snap frozen in liquid nitrogen and homogenized in a blender (Terzano, Milan, Italy). Bovine inner ear powder, bovine spleen powder and bovine liver powder (Sigma, St Louis, MO, USA) were resuspended in 50 mm TRIS buffer (pH 6·8), 20 mm EDTA, 2 mm phenylmethylsulphonylfluoride (PMSF) and 10 µg/ml of leupeptin, then sonicated in an ice bath and centrifuged at 20 000 g for 15 min at 4°C. The protein concentration of the supernatant was determined by the BCA method (Pierce Chemical Company).

Western blot analysis

The extracts were separated by 10% SDS-PAGE under reducing conditions using 20 µg protein/mm gel, and transferred to nitrocellulose filter paper [10]. The nitrocellulose filters were then cut into 0·5 cm strips and saturated for a 1-h incubation period in 0·05 m Tris, 0·15 m NaCl, and 5% dry non-fat milk.

Sera diluted in Tris buffered saline, 5% fetal calf serum, 0·05% Tween at a concentration of 1 : 250 were incubated on the filters for 3 h at room temperature. After repeated washings, the bound antibodies were detected by goat anti-human IgG F(ab′)₂ conjugated with alkaline phosphatase (Sigma). The immunoactive bands were visualized using 5-bromo-4-chloro-indolyl-phosphate (BCIP) and nitroblue tetrazolium (NBT) as substrates [11].

Antibodies to nonorgan specific antigens

Anti-nuclear antibodies were detected by immunofluorescence on Hep2 cells. Antibodies to extractable nuclear antigen (ENA) were detected by counterimmunoelectrophoresis and anti-neutrophilic cytoplasmic antibodies (ANCA) by ELISA using myeloperoxidase and proteinase 3 as antigens.

Statistical analysis

Statistical analysis was carried out to detect correlations between the presence of antibodies to specific inner ear antigens and the demographic and/or clinical features of the patients. Only antigens detected by MD patients and not by normal sera were taken into account. The Student's t-test, χ² test, Fisher's exact test and multiple logistic regression test were used as appropriate.

RESULTS

Twenty-five sera from MD patients and 38 from healthy controls were tested by immunoblotting on inner ear extracts. Normal sera reacted with several antigens over a wide range of molecular weights. Most of the MD sera (23/25) also reacted with additional antigens that were not detected by normal sera, whose molecular weights ranged between 128 and 42 kD. A representative blot showing the reactivity of MD sera with inner ear extract is presented in Fig. 1. A small number of antigens were detected by a high proportion of MD sera (but not by normal sera): 11/25 (44%) patients showed reactivity to a protein of 44 kD and 10/25 (40%) to a protein of 53 kD. Eight patients displayed reactivity to both the 44 and 53 kD antigens. Coexpression of the two autoantibodies is significantly higher than the single reactivity (P = 0·005 by Fisher's exact test).

Fig. 1.

Reactivity of the sera from Meniere's disease patients and healthy subjects on inner ear extract. Nitrocellulose strips nos. 1 and 2 were incubated with sera from healthy donors and nos. 3, 4, 5, 6 with sera from patients with Meniere's disease. No antigens are detected by the sera of healthy subjects, while sera from the patients detect antigens of different molecular weight, in particular 53 kD (strips 3, 4, 6) and 44 kD antigens (strips 4, 5, 6).

To evaluate the organ specificity of the antibodies that reacted with inner ear antigens, sera from patients and controls were also tested on bovine liver and spleen tissue extracts. Only 2 out of 25 patient sera reacted with a 44 kD protein on spleen and 3/25 with a 44 kD protein on liver, while 2/25 sera showed reactivity to a 53 kD antigen on spleen and 1/25 sera with a protein of the same apparent molecular weight on liver. A representative example of the reactivity of MD sera with inner ear, spleen and liver extracts is shown in Fig. 2. Thus, the immune response to the 44 kD and the 53 kD proteins seems to be disease-specific, as the antibodies to these antigens are present only in MD. Moreover, since proteins with the same apparent molecular weight are rarely detected in other tissues, the 44 and 53 kD antigens can be considered inner-ear specific antigens. Adjacent strips containing spleen, liver and ear extracts were probed with a monoclonal anti-β actin antibody and with serum from a MD patient reacting with a 44 kD antigen in all the extracts (Fig. 3). The 44 kD antigen can easily be distinguished from β actin.

Fig. 2.

Reactivity of sera from four Meniere's disease patients (patient nos. 3, 4, 5 and 6) on inner ear, spleen and liver extracts. E, bovine inner ear extract; L, bovine liver extract; S, bovine spleen extract Reactivity to 44 and 53 kD antigens are detected on inner ear extract, but not on spleen and liver extracts.

Fig. 3.

Reactivity on spleen (S), liver (L) and ear (E) extracts of monoclonal anti-β actin antibody and serum of a MD patient reacting with 44 kD antigen. The 44 kD antigen can easily be distinguished from β actin.

Among the non-organ specific antibodies, anti-nuclear antibodies at low titre (1 : 40) were detected in 4/25 (16%) patients; antibodies to extractable nuclear antigen (ENA) or anti-neutrophilic cytoplasmic antibodies (ANCA) were not detected in any serum.

We then sought to correlate the specificity of these autoantibodies with the clinical features of the disease. Reactivity with the 44 or 53 kD proteins was not correlated to age, sex or disease duration. Analogously, the levels and types of hearing loss were not significantly different between patients who were positive or negative for these autoantibodies: the PTA values were indeed similar (52·4 ± 13·4 dB versus 51·4 ± 12·4 dB) and the type of hearing loss did not differ (6 FHL and 8 LFHL versus 8 FHL and 3 LFH)

Finally, using the multiple logistic regression test we did not detect any correlation between the presence of these autoantibodies and the clinical features of the disease globally considered.

DISCUSSION

Autoimmune inner ear disease was first described by McCabe in 1979 [12] as sensorineural hearing loss (SNHL). The clinical presentation of SNHL can be quite variable, often overlapping with other disorders such as MD. Hughes et al. [13] reported that over 52% of patients diagnosed with autoimmune inner ear disease presented with hearing loss and true vertigo. This suggests that a continuum may exist between MD and SNHL. The pathogenesis of MD is still unknown. However, it has been proposed that viral infections and autoimmune processes may play a role in the induction of the disease. The immune response in patients affected by autoimmune inner ear disease has been assessed in a number of studies using different techniques: lymphocyte transformation tests [14]; measuring the levels of anti-CII and anti-laminin antibodies on ELISA [4–7]; indirect immunofluorescence on ear sections [15,16]; and more recently Western blotting[17–20].

In 1990 Harris and Sharp [17] reported that a significant proportion of sera from patients affected by SNHL, when analysed on Western blots against fresh bovine inner ear extracts, showed antibodies to a 68 kD protein. The presence of the antibody was correlated with disease activity and the response to steroid therapy. Antibodies to a 68 kD antigen were also detected in two series of Meniere's disease patients at a frequency of 47% and 33%, respectively [21,22]. Gottschlich et al. [23] similarly found that a significant proportion of patients affected by MD or bilateral, rapidly progressive sensorineural hearing loss showed reactivity to a 68 kD protein, and suggested that this antibody might be used as a diagnostic marker. The protein was subsequently characterized in greater detail and found to be ubiquitous rather than inner ear-specific, and eventually was identified as hsp70 [24]. Patients with autoimmune inner ear disease frequently react with an other ubiquitous protein of 42 kD highly conserved among different species that was identified as β actin [25].

In this disorder, however, autoantibodies reactive with inner-ear specific antigens have also been detected. Yamanobe and Harris [18] demonstrated that sera from patients with SNHL react with ear-specific proteins of 32 and 35 kD and Gottschlich et al. [23] also report, besides the frequent reactivity with the 68 kD antigen, the presence of antibodies specific for several ear proteins including a 58 kD antigen. This 58 kD protein was more extensively studied by Boulassel et al. [26] that identified it by direct sequencing as the cochlear antigen COCH5B2.

We previously studied the levels of antibodies to CII, CV and CI and to basement membrane proteins (laminin and CIV) in our MD patients [7]. A low incidence of anticollagen and anti-laminin antibodies was found. In particular, anti-CII antibodies were present in only 15·5% of the patients. The results of our study did not indicate a clear role of autoimmune responses to basal membrane antigens and collagen in the pathogenesis of MD.

We therefore decided to study the reactivity to bovine ear extracts of sera from patients previously tested for anti-collagen and anti-laminin antibodies. No patient showed reactivity to the 68 kD protein, whereas 11/25 sera reacted to a protein of 44 kD and 10/25 to a protein of 53 kD. When we tested the same sera on bovine liver and spleen the percentage of reactivity to the 44 and 53 kD antigens was very low, suggesting that these proteins are not ubiquitous.

Our results are consistent with previous reports on the reactivity of sera from patients with autoimmune inner ear diseases. As reviewed by Yoo et al. [6], the antigens most commonly detected by these patients’ sera are in fact 32–35 kD, 42–46 kD, 52–58 kD, 57–68 kD and 79–80 kD. Antigens of similar molecular weight have been detected in extracts from different species, suggesting once again that organ-specific and not species-specific antigens are the target of autoantibodies. On the contrary, antibodies to ubiquitous antigens such as basement membrane proteins and nuclear antigens, which are commonly recognized by systemic autoimmune disease sera, are not present in Meniere's disease patients. At the same time, signs and symptoms of autoimmune disorders are not present in these patients and significant associations with connective tissue disorders have not been found.

In the attempt to identify MD-specific autoantibodies, we selected for our study a highly homogeneous population of 25 patients, with monolateral disease of more than 9 years duration. The non complete overlap between our results and the ones reported in the literature can be in part explained by these restrictive inclusion criteria. The procedure used to obtain the ear extract should be also taken into account. In fact, we used a low salt extraction, that may be inefficient to solubilize membrane proteins thus limiting the detection of some autoantibody specificity.

However, despite differences in patient selection and in experimental design, the results obtained in many studies, including ours, suggest that in MD the immune response is focused on inner ear antigens. It remains unclear whether the antibodies to 44 and 53 kD proteins play a role in the pathogenesis of MD or if they represent the result of inflammation and tissue destruction. Even if the latter is the case, they could contribute to the perpetuation of the disease or play a role as cofactors in association with other mechanisms. Analysis of patients in the early phase of the disease might help to clarify the pathogenic role of these autoantibodies.

A study is currently underway to confirm our results in a larger group of patients, as the data reported here was obtained in a limited number of subjects, albeit with a carefully designed protocol and a long follow-up period. Western blotting is the first step in antigen identification: in fact, the molecular weight is only one of the many features that define a protein and a number of antigens from the same tissue may share an identical molecular weight. Thus, only the biochemical characterization of these inner ear autoantigens can shed more light on the pathogenetic mechanisms of MD.